Upload
others
View
6
Download
0
Embed Size (px)
Citation preview
NICER, Gravitational Waves, and Neutron Stars
M. Coleman MillerUniversity of Maryland, Astronomy Department
Joint Space-Science Institute
Outline
• The importance of neutron star radii• NICER measurements of mass and radius
of PSR J0030+0451
Will talk only about our work (Miller, Lamb, Dittmann+ 2019)Please also read other papers in the ApJ Letters focus issue,especially Riley et al. 2019; Raaijmakers et al. 2019;Bilous et al. 2019Key point: favored models from the two NICER groups are fully consistent with each other in M, R, and spot patterns
Questions During Talk
• Please feel free to ask questions at any time
• I will also pause twice during the talk to determine whether anyone would like to pursue discussion points
But First: The Main Results• For the 205.53 Hz pulsar PSR J0030+0451
Isolated pulsar: no indep knowledge of M• Equatorial radius• Gravitational mass• Best configuration has three spots; almost
equally good configuration has two spots• All spots are in the rotational hemisphere
opposite observer. At least one spot is highly elongated
The Importance of Radii• Radius would provide
great EOS leverageWide range in models
• But tough to measure• Previous published
measurements are susceptible to huge systematic error
• NICER X-ray pulse modeling can help Demorest+ 2010
The Importance of Radii• Radius would provide
great EOS leverageWide range in models
• But tough to measure• Previous published
measurements are susceptible to huge systematic error
• NICER X-ray pulse modeling can help Demorest+ 2010
Radius Bias with T VariationExample of the bias toward low radii from single-tempfits to surface with varying temperature.
Temperature varies smoothly from 2 keV(equator) to 0.2 keV (pole).
Fit is good, but R is 13% low. With narrower T profile, correction is larger
Good fit and lack of pulsations does notguarantee uniformity!
Assume perfect energy response, zero NH
Key: Minimal Systematic Errors
• Extensive work by Fred Lamb (Illinois) and myself with our collaborators suggests that when we fit energy-dependent waveforms, systematic errors are minimized
• We have generated synthetic data using models with different beaming, spectra, spot shapes, temperature distributions etc. than used in fitting the data
• Conclusion: if good fit, no significant bias
The Idea in Brief
Bayesian fits: trace rays from hot spots on NSsurface, compare with energy-dep waveform
Concern about rotation?
• Fundamentally, we are tracing photons from the star to the observer
• If star is not rotating, this is relatively simple: no rotation means spherical symmetry, so a given photon travels in a plane
• Not true when there is rotation; frame-dragging.
• Also, star becomes oblate
Frame-dragging doesn’t matter
3
4
5
6
7
8
9
f E (
10-4
cm
-2 s-
1 keV
-1)
NumericalS+D
OS
-0.5
0
0.5
1
1.5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Frac
tiona
l diff
. (%
)
Rotational Phase
S+D DifferenceOS Difference
Shape DifferenceNumerical Error
Approximations:S+D: star is spherical,Schwarzschild+SR raytracing.OS: star is oblate,Schwarzschild+SR raytracing.Compare with fullnumerical waveform
Conclusion: to theprecision we need, wecan treat spacetime as if there is no rotation
Bogdanov et al. 2019
Figure by Sharon Morsink based on originalconcept by Scott Lawrence (UMd)
n=200 Hz
Effect of Rotation on M-R Curves
M vs. R for four EOS, at 200 Hz vs. 0 Hz. Difference is negligible compared with measurement precision.Calculations by Sharon Morsink.
Models Used in Fits
• We consider uniform-temperature spotsPossibly different T; arbitrary locations
• Each spot can be oval: start with a circular spot and stretch or squash it azimuthallyFits include unmodulated background
• Fits use two or three oval spotsArbitrary overlap of spotsGives great flexibility of modeling (e.g., can have isolated spots, or crescents)
Fit to Synthetic Two-oval Data
Inner contour: 68% of posterior probabilityOuter contour: 95% of posterior probability
Any Questions At This Stage?
Mass-Radius Posteriors for J0030
Left: M-R posterior for NICER J0030 data, two ovalsRight: M-R posterior for NICER J0030 data, three ovals
1D Posteriors: NICER 2,3-oval
Top: analysis of NICER data, two-oval modelBottom: analysis of NICER data, three-oval modelDotted line on right: distance prior
Gaussian prior on distance: ; chan 40-299
Bolometric Waveforms
Left: two-oval model fits to NICER J0030 dataRight: three-oval model fits to NICER J0030 dataDotted lines are individual spots; solid, total
Bolometric Residuals
Residuals of best-fit three-oval model comparedwith J0030 NICER data, for 64 phases. Fit is good
Phase-Channel Residuals
Residuals (in c) for best three-oval fit to NICERJ0030 data. No patterns are evident, as one wouldexpect from a good fit (c2/dof=8189/8040, 12%)
Spot Patterns
Top: two-oval fit. Bottom: three-oval fitHorizontal solid line shows observer inclination
Shouldn’t B be a centered dipole?
• Uranus’ and Neptune’s fields aren’t!• Millisecond pulsars go through complex
evolution; B, spots need not be simple
Credit: NASA
Any Questions?
NICER Contribution to EOSRed line: ratio of the5%-95% pressure rangewhen NICER (M,R) from J0030 is included, to therange prior to NICER, as afunction of density
NICER M and R reducespressure range by 10-30%from ~rsat to 2rsat
Exposure time will ~doubleby end of 2020.Can incorporate into full EOS
constraints: Miller, Chirenti, Lamb 2020, many other papers
Implications for Equation of State
Top: spectral EOS. Bottom: piecewise polytropeLeft: prior (dot-dash 0%-100%; solid 5%-95%)Middle: result of adding NICER M-R for J0030; 5%-95%Right: result of also adding high-M and L upper limit
Dashed lines:Hebeler+ EOS
Conclusions• First NICER measurements, for PSR
J0030+0451, have already tightened EOS constraints. Full, (M,R) posterior samples:https://zenodo.org/record/3473466
• Key: measurements appear reliable as well as precise
• Doubling+ of data set and contributions from analysis of other pulsars (especially J0437 [best precision] and J0740 [highest mass]) will improve constraints substantially